When AI Goes After The Pigeons

There are plenty of sophisticated uses for computer vision and automated systems. Then there is someone who got annoyed with pigeons and built an automated water cannon to deal with them.

The homemade system uses a camera and computer vision to recognise pigeons when they appear, before automatically aiming a water gun in their direction and firing. It’s an entertaining solution to a very ordinary problem, but underneath it is a good demonstration of how accessible some fairly sophisticated technology has become.

None of the individual components are particularly new, but they are becoming much easier and cheaper to combine. Open-source models, affordable cameras and small computers mean you don’t necessarily need a research lab or huge budget to start experimenting with systems that can recognise what’s happening around them and respond.

In this case, the response happens to be a wet pigeon.

From General Technology To Personal Automation

Much of the conversation around AI automation focuses on organisations looking for processes they can automate at scale, where the time or money saved helps justify the investment. Personal automation works rather differently because there doesn’t need to be a business case or a huge potential market; somebody can simply have a problem that annoys them enough to wonder whether they can build something to solve it.

That’s what I find particularly interesting about the pigeon cannon. It’s highly specific, probably completely unnecessary for the vast majority of people and doesn’t need to become a commercial product to be useful to the person who made it.

As the technology becomes cheaper and easier to experiment with, we could see much more of this kind of personal automation. People can create systems around their homes, hobbies and everyday frustrations, even when what they build would be completely pointless to almost everybody else.

It also changes where new uses for AI might come from. We tend to hear about applications developed by technology companies and then offered to consumers, but accessible tools give individuals much more freedom to start with their own problem and work backwards from there.

Pigeons Today…

The slightly ridiculous thing about an autonomous pigeon water cannon is that the basic structure behind it isn’t particularly ridiculous at all. A camera observes an environment, AI identifies something within it and, if certain conditions are met, a physical action follows.

That same pattern has applications across security, traffic management, warehouses, agriculture and robotics. The pigeon cannon is simply a particularly entertaining example of how much easier those pieces are becoming to put together.

And that may be what makes experiments like this worth paying attention to. New technology doesn’t only become significant when somebody finds a huge commercial application for it; accessibility matters too, because it allows far more people to experiment with what the technology can do and decide what problems they want to use it to solve.